|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- audiofolder |
|
- lucas-meyer/asr_af |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: xls-r-asr_af-run4 |
|
results: |
|
- task: |
|
name: Automatic Speech Recognition |
|
type: automatic-speech-recognition |
|
dataset: |
|
name: audiofolder |
|
type: audiofolder |
|
config: default |
|
split: validation |
|
args: default |
|
metrics: |
|
- name: Wer |
|
type: wer |
|
value: 0.36875288328463784 |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# xls-r-asr_af-run4 |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the asr_af dataset. |
|
It achieves the following results: |
|
- Wer (Validation): 37.89% |
|
- Wer (Test): 38.41% |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 0.0003 |
|
- train_batch_size: 4 |
|
- eval_batch_size: 4 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 3 |
|
- total_train_batch_size: 12 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 30 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer (Train) | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:| |
|
| 4.1702 | 1.76 | 400 | 1.1378 | 0.8201 | |
|
| 0.6633 | 3.52 | 800 | 0.5165 | 0.4819 | |
|
| 0.3114 | 5.29 | 1200 | 0.4763 | 0.4115 | |
|
| 0.1986 | 7.05 | 1600 | 0.5097 | 0.3923 | |
|
| 0.136 | 8.81 | 2000 | 0.4876 | 0.3829 | |
|
| 0.1098 | 10.57 | 2400 | 0.5036 | 0.3688 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.28.0 |
|
- Pytorch 2.0.1+cu118 |
|
- Datasets 2.14.4 |
|
- Tokenizers 0.13.3 |
